There’s a gospel choir on the record I’m releasing this month. Consider this podcast episode the release party. Sixteen voices, brass underneath, a room you can hear the size of. It comes in about two-thirds of the way through a song called “F**k That Family,” which is about the people who chose you back — the ones who aren’t blood and set a place for you anyway.

That choir doesn’t exist.

I typed a description of what I wanted at my kitchen table in Barcelona at four in the morning, and waited about ninety seconds.

I’ve spent years on the other side of the glass, producing music videos for working regional acts. I know what a record costs at that level. A sixteen-piece choir with a horn section is a five-figure day before anyone has mixed anything. No regional act in Europe can afford it. Not one.

I didn’t hire anything.

This week’s episode is about what happens next. And I should say at the top what I say at the top of the episode: I’m not a neutral party in this. Normally on this show I’m at some distance from what I’m covering — somebody built a thing, somebody sold a thing, somebody got displaced, and I go find out what happened and why it matters. Not this time. This time I’m the one who did it.

I’m putting myself on trial here.

What I made

Eight songs. About forty hours. It’s called Suburb City, by 50 Minute Hour.

It’s a concept album — one narrator, multiple voices, one town, eight chapters. It’s heavy on Gen-X angst, anthems, and catharsis. And it is very personal to me, made with love and care — by me and the machine. That’s the contradiction.

There’s a Music Man patter parody about the therapy industry, sung by a huckster selling wounds instead of pool tables. A waltz in three-four where every argument resolves into you misremembered. An orchestral lament about three attempts to leave a marriage, each one interrupted. A gospel showstopper. A rockabilly strut about resigning from the position of designated family failure. Glam rock. Pop punk. And a rock opera anthem set in a tower where everyone who ever damaged you lives on a different floor, and you’re the one paying the heating bill.

Every word is mine. Every arrangement decision is mine. I knew exactly what I wanted each song to do and I iterated until it did it.

And every single performance on that record was generated. All of it. There is not one human being on it besides me, and I don’t sing.

It was made with love, passion, and commitment — and it would not exist otherwise. No budget, no band, no time. I’ve wanted to write music since I was fifteen, which is also the year I won a national poetry prize for a parody of Shakespeare. Act Two, Scene Seven of As You Like It. The one about all the world being a stage.

What it replaced

Here’s the part I haven’t heard anyone in my position say out loud.

That record replaced a session drummer.

It replaced a bass player. A guitarist. A horn section. A sixteen-member gospel choir. A tracking engineer. A mix engineer. A mastering engineer.

Nine or ten paid jobs. Forty hours. Zero pounds, euros, or dollars.

I want to be precise about the verb, because it matters. I didn’t evaluate those people and decide I could do without them. I didn’t audition anyone and find them wanting.

I couldn’t. I skipped them, because skipping them was available.

That’s the real story of the last three years, and almost nobody is telling it straight — least of all people like me.

The wrong question

In November of last year, Deezer and Ipsos ran the biggest study anyone’s run on this. Nine thousand people, eight countries. They played them music and asked whether it was made by a human or a machine.

Ninety-seven percent got it wrong.

That number has been in every article about this for eight months. It’s in open letters. It’s in policy submissions. It gets cited by the people worried about all this as proof of a crisis, and by the people building these tools as proof of arrival.

Both sides love it, which should be your first clue that something’s off.

Because here’s what that study actually measured: detection. Can you tell?

And I’d ask you — when have you ever chosen music that way?

Think about the last song you loved. Did you love it because you correctly identified the production chain? Did you put it on and think, ah yes, session players, analogue desk, human hands? Or was there autotune on it? Pitch correction? Stem separation? AI plug-ins in Logic or Pro Tools or Ableton that plenty of musicians use and nobody discloses?

The questions people actually ask are much older and much simpler. Did you like it. Was it any good. Did whoever made this work hard at it.

Nobody has tested any of those. Not once.

Every blind test that’s been run compares different songs, by different writers, and asks people to spot the machine. Which leaves the biggest variable in music — the writing — completely loose. If a listener prefers the human track, was that the performance? Or did they just prefer the song?

Eight months of argument, thousands of pages of policy, platform bans, lawsuits. And the entire evidentiary basis is a study that answered a question nobody asks.

The books on the shelf

I’ve watched this happen before. Several times.

I started using Photoshop at version one, in 1989, before it had layers. Then PageMaker, QuarkXPress, Freehand, Illustrator, InDesign. I watched nonlinear editing arrive and empty out the flatbed rooms — rooms full of people who cut film with their hands, and then those rooms were storage.

I watched desktop publishing kill typesetting. I used Linotronic typesetters, wax, and brayers in print shops when I was fifteen, doing paste-ups for my high school paper at the Landrum News Leader in Landrum, South Carolina. An entire skilled trade, with apprenticeships and specialists and decades of accumulated knowledge, gone in about ten years.

Nobody voted on that. Nobody consulted the typesetters.

But here’s the part of that history I’ve never heard anyone bring into this argument.

The design annuals.

If you worked in a corporate graphics department any time between about 1975 and 2005, you know exactly what I mean. Every January the department ordered the books. Communication Arts. Print. The Type Directors Club annual. The AIGA annuals. Beautiful, hardbound, ninety dollars a pop, full of the year’s award-winning work.

Nobody bought them to be inspired.

You bought them so that when a client needed a quarterly newsletter by Thursday, your in-house designer could open to page eighty-four, find the layout that won a gold medal in Chicago, and rebuild it with the client’s logo on it.

That was the job. That was how in-house design got made, at scale, across an entire industry, for thirty years. The award-winning designers knew. Everybody knew. The annuals kept selling.

So let’s call it what it was: a training corpus with a spine and a dust jacket and a purchase order attached. And we called it professional development, and nobody in forty years called it theft.

Now — here’s where the analogy breaks, and I don’t skip past it in the episode, because if I did the whole thing would be a cheap shot.

Scale and consent.

The annuals were a transaction. Somebody got paid. The designers submitted their work knowing what the book was and what it was for. Ninety dollars changed hands and everybody understood the arrangement.

What happened with these models is neither of those things. Nobody submitted. Nobody was paid. The scale isn’t a bookshelf — it’s most of recorded human output.

That difference is real and it’s the strongest argument the other side has. I’m not waving it away.

What Bowie saw

In June of 2002, David Bowie told the New York Times that music was about to become like running water, or electricity. A utility. That copyright would no longer exist in ten years, and the only unique situation left would be showing up and playing it in a room. You’d better be prepared, he said, for doing a lot of touring.

That’s four years before Spotify was founded.

Everybody treated it as the eccentric talking.

He was mostly right, and very early. Copyright didn’t die — he got that part wrong, and I’ll own it on his behalf. But the utility, and the touring, and the collapse of the recording as the thing you sell: all of that happened.

And he’d already acted on it. In 1997 he raised fifty-five million dollars against his own future royalties — twenty-five albums, two hundred and eighty-seven songs, everything before 1990. Took the money up front. Used it to buy back his masters from his former manager.

Napster launched two years later. By 2004, Moody’s had cut the rating to one notch above junk.

He wasn’t predicting AI. He was predicting something more useful: that the recording was never going to be the valuable part.

The room was.

The part nobody’s campaigning about

Almost all of the public argument about AI and music is about voices. That’s not an accident — vocals are the part a non-musician can point at, and there’s a real emotional charge to a synthetic person singing to you about their feelings.

But the work that’s actually disappearing is quieter and further down.

It’s the second guitarist who came up with the part that made the chorus work. It’s the drummer who played it four times and the fourth one was the one. It’s the engineer who’s been mixing for twenty years and knows, without thinking about it, that the second chorus needs the snare pulled back four milliseconds or the whole thing sits wrong.

None of that is legible to a listener. All of it is why records sound like records.

There’s no open letter about the person who knew which room to book.

So here’s what I want to do about it

The album is licensed Creative Commons, attribution only. All eight songs. Free.

Any musician, any band, any producer who wants to take one of these songs and rebuild it with human beings — players, singers, tracking, mixing, mastering, the whole chain — it’s yours.

The words are mine and the structure is mine. Everything else is negotiable. Change the melody. Change the tempo. Change the genre entirely. Turn the waltz into a dirge. Turn the gospel number into a punk song. I’m not precious about anything except the words, and honestly I’ll argue about those if you make a good case.

Credit me as the writer and it’s yours. Record it, release it, sell it, play it live. I’m not asking permission back.

Then both versions go up. Same page, same player, level-matched, unlabeled.

And listeners answer one question. Not can you tell.

Which one would you play again?

That’s it. That’s the whole idea. And the reason it’s worth doing is that it controls for the thing every previous test left wide open — same words, same songs, same writer. The only variable that changes is who, or what, performed it.

Whatever difference shows up is the value of human musicianship, measured. Not asserted in an open letter. Measured.

And whatever comes back, I publish. If the human version is better, that goes up. If it isn’t, that goes up too.

Why I actually care

There’s a song on this record about a woman who died. She was my wife for a long time, and then she wasn’t, and then eighteen months after I finally got out, the phone rang.

There’s another song with a spoken passage in the middle of it. It’s a sentence somebody said to me across a kitchen table, more or less word for word, and I put it in because it was true and because I’d never been able to explain it to anyone otherwise.

The machine sings both of those beautifully. Pitch is perfect. Phrasing is right. Every note where it should be.

It doesn’t mean it.

And I honestly don’t know whether that’s a real thing I’m hearing, or whether it’s sentiment — whether I just know what’s underneath those songs and I’m supplying the weight myself.

Nobody knows. That’s the whole point. Nobody has checked.

Bowie said the room was the thing. I don’t have a room.

The songs are yours — Creative Commons, attribution only.

I’d like to hear somebody mean it.

🎧 Listen to the full episode above.

🎵 The album is here: www.citizenone.world/suburbcity

Citizen One — Exploring Our Urban Future.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit multiversethinking.substack.com